Daniel Bloom, a PM at Melio (fintech), shows Claire Vo his Claude Cowork system on How I AI — a personal agent harness that manages his week: “I’m able to do in a day what used to take me a week.”
The two rules of a powerful system
- It can rewrite its own core files — the system keeps improving itself
- It connects to as much of your ecosystem as possible — Notion, calendar, Slack, Gmail, Granola meeting transcripts
- The tool matters less than these two properties — Cowork works for him, but Codex or ChatGPT Work could do the same
The architecture
- Notion as a read-only brain: three columns (Top of Mind / This Week / Inbox) — Cowork built the board itself when it got tired of his messy Google Doc, and manages it on his behalf
- Context files: a CLAUDE.md-style context file for every work area, goal, and colleague; he spent the first weeks “contextualizing ruthlessly” — feeding links, decks, and endless voice notes (Whisper)
- Weekly prep (Sunday): a recurring task composed of skills — pulls his whole ecosystem, recommends the week’s focus, triages meetings into serious-prep / quick-reminder / nothing
- Daily brief (morning): walks through yesterday’s meetings via Granola transcripts with one-liners and action items, then asks what to expand or draft
The killer feature: proactive context
The daily brief scans recent Slack/email/notes for things it doesn’t understand — new files, milestones, goals, internal terms — and asks him to define them, then saves them to context. Internal vocabulary like “settlement cap” never makes it into training data; this is how the system learns the company’s real language. Claire’s verdict: “really sharp, something we haven’t seen on the podcast.”
It also knows his workflow: inbox-zero on Slack and email means a saved item that’s no longer saved is probably done.
The self-improvement loop (weekly, scheduled)
- Draft-gap learning: finds drafts Cowork wrote that he rewrote before sending, and learns the delta — his actual writing style
- Skill suggestions: spots recurring actions and proposes turning them into skills
- Feedback telemetry: every skill collects friction lines from his interactions; weekly review surfaces the top issues
- The “improve” skill: a critical auditor for the endless flood of AI-setup tips — he dumps links into a Slack channel and Claude judges “is this real? does this fit my system?” before they decide together whether to build it
Scaling to the team: the Workstation
- Melio’s onboarding plugin: a guided chat UX that connects your tools, maps roles/colleagues/calendar, clones your writing voice, and captures goals — productive from day zero in ~15 minutes
- Built after watching PMs spend three days installing a Claude Code terminal setup; the lesson: a tool tuned to how you work fails for everyone else until it has a guided UX
The missing piece
- ~70-80% of his computer time runs through Cowork; the gap is cloud-resident operation — the agent acting when his laptop is off (he names Claude Tag, OpenClaw, and Hermes as the direction)
- He’s preparing for it: Cowork already scores every captured task with what it imagines a closed scenario looks like, so the simplest asks (file requests, scheduling, replies) can eventually resolve on their own
“I’m really able to do in a day now what used to take me a week — and more than that, it’s deeper work. There are things I do today that I just wasn’t able to do before.”